Lake origin determines Daphnia
population growth under winter
conditions
CHRISTIAN RELLSTAB1,2*†
AND PIET SPAAK1,2 1
EAWAG,SWISS FEDERAL INSTITUTE OF AQUATIC SCIENCE AND TECHNOLOGY CH-8600,DU¨ BENDORF,SWITZERLAND AND2INSTITUTE OF INTEGRATIVE BIOLOGY, ETH ZU¨ RICH,CH-8092ZU¨ RICH,SWITZERLAND
†
PRESENT ADDRESS:DEPARTMENT OF BIOLOGICAL AND ENVIRONMENTAL SCIENCE,CENTRE OF EXCELLENCE IN EVOLUTIONARY RESEARCH,UNIVERSITY OF JYVA¨ SKYLA¨,PO BOX35,FI-40014,FINLAND.
*CORRESPONDING AUTHOR: [email protected]
Received September 22, 2008; accepted in principle November 14, 2008; accepted for publication November 19, 2008; published online 23 December, 2008
Corresponding editor: John Dolan
In large oligotrophic lakes, growth rates of zooplankton populations decline to low or negative values in winter as a result of low food concentration and water temperature. Daphnia, a key species in the aquatic food web, has two strategies to overwinter: either by sexual reproduction resulting in diapause or as asexual clones in the open water. We investigated how asexually overwintering Daphnia clones, originating from different taxa and lakes with different trophic state, survive under winter conditions. We performed a laboratory experiment exposing D. hyalina and D. hyalinagaleata clones to low water temperature (58C) and either no or low food (,0.1 mg C L21) conditions. We used clones from three pre-alpine lakes in Switzerland with contrasting trophic state: ultra-oligotrophic Lake Brienz, mesotrophic Lake Constance and eutrophic Greifensee. Our results show that Daphnia clones can withstand starvation for an average of 6 weeks under low food concentration and for almost 2 weeks under complete absence of food. Besides strong clonal variation, we found a significant effect of food concentration and its inter-action with lake origin on population growth rate. Our study indicates that adaptation to local winter conditions is a key factor in defining the clonal, but not necessarily the taxonomic compo-sition of a Daphnia population in unproductive lakes.
I N T RO D U C T I O N
The cladoceran Daphnia, a key species in the aquatic food web, has two overwintering strategies (De Meester et al., 2006). One strategy involves sexually produced diapausing eggs that hatch when conditions ameliorate, for example, in the following spring, and another involves asexual females that survive the winter in open water. Although permanent large lakes favour asexual reproduction (Mort and Wolf, 1986), often both modes of reproduction are found together (Jankowski and Straile, 2004; Keller and Spaak, 2004). Sexual reproduc-tion forms new genotypes through recombinareproduc-tion, but is energetically costly (Lynch et al., 1986) and requires the
temporal and spatial co-occurrence of males and sexual females (Keller et al., 2007). Overwintering asexual clones have to cope with strong selection factors: low water temperature and, at least in unproductive systems, low food concentration. In oligotrophic lakes, food con-centration can drop below food threshold levels (Lampert, 1977, 1978; Gliwicz, 1990) for several con-secutive months (Rellstab and Spaak, 2007). If overwin-tering asexual clones survive this period of starvation, they could have a competitive advantage over sexually produced clones in spring when conditions ameliorate, because they are already present in the water column and do not depend on hatching cues. Variation in the ability of clones and taxa to survive periods of starvation
under low water temperature conditions can thus affect the clonal and taxonomic composition of Daphnia popu-lations and influence the dynamics of the whole food web. Several studies have been made of the effect of star-vation on Daphnia (Elendt, 1989; Gliwicz, 1991; Epp, 1996), and many of them focus on short-time starvation (Plath, 1998) and subsequent refeeding (Bradley et al., 1991). Resistance of daphnids to starvation depends on the amount of energy reserves, i.e. lipids (Lemcke and Lampert, 1975) and the allocation of this energy into reproduction or survival (Tessier et al., 1983; Bradley et al., 1991; McCauley and Nisbet, 1991). In general, smaller species and individuals show lower survival under starvation than larger ones (Gliwicz, 1991), and juveniles show lower survival than adults (Threlkeld, 1976). This is consistent with the size-efficiency hypoth-esis which states that larger bodied species or individ-uals should be superior, as the feeding rate increases faster than the respiration rate with increasing body size (Brooks and Dodson, 1965). Maternal effects (Mousseau and Fox, 1998) play an important role in the ability to survive starvation, at least in early life stages (Lynch and Ennis, 1983), because lipids are passed from the mother to her offspring (Tessier et al., 1983). Starvation resist-ance can differ among Daphnia species (Threlkeld, 1976), but also intraspecific clonal variation has been shown (Epp, 1996).
Although periods of low food levels often occur in winter in pre-alpine lakes, none of the above-mentioned studies used water temperatures typical for that season. Temperature is an important factor influencing life-histories of Daphnia because it changes, for example, the processes of energy intake, assimilation and respiration losses (Burns, 1968; Yurista, 1999). Studies testing temp-eratures experienced in natural environments during winter are rare (Bottrell, 1975; Vijverberg, 1980; Goss and Bunting, 1983). Rising temperatures generally speed up rates of development or growth by decreasing par-ameters such as instar duration (Vijverberg, 1980), number of instars to first reproduction (Goss and Bunting, 1983) or egg development time (Bottrell, 1975), and increasing the clutch size (Goss and Bunting, 1983) and the growth rate (Vijverberg, 1980). Often these factors show an optimal response temperature which can be between 15 and 258C depending on the Daphnia species. To our knowledge, no study exists that investi-gates the effect of temperature experienced by Daphnia during winter on its lifespan under constant food con-centration, but it is generally assumed that the low temperature increases lifespan as a consequence of a lower “rate of living” (Pearl, 1928), as discussed in Lynch and Ennis, and Gliwicz et al. (Lynch and Ennis, 1983; Gliwicz et al., 2001).
The pre-alpine lakes of Switzerland differ greatly in their trophic state and therefore provide an ideal oppor-tunity to study the effect of low food conditions, or the complete absence of food, on the performance of different Daphnia taxa. Large lakes with low phosphorus content are dominated by D. hyalina, whereas D. galeata prefers more productive habitats. Daphnia hyalina galeata hybrids are as common in Swiss lakes as in other European lakes (Schwenk and Spaak, 1995) and most frequent in lakes with the largest changes in magnitude of phosphorus content since the peak of eutrophication in the 1970s (Keller et al., 2008). Daphnia hyalina has the tendency to overwinter asexually, whereas D. galeata relies more on resting eggs (Jankowski and Straile, 2004).
In the present study, we wanted to investigate how various clones from the D. hyalina/galeata species complex cope with low food concentration and the low water temperatures typical for winter in pre-alpine oli-gotrophic lakes. In particular, we were interested in the survival and reproduction of Daphnia under these cir-cumstances. We hypothesized that (i) high mortality and no reproduction would be found at extremely low food conditions; (ii) D. hyalina clones would be more resistant to starvation than D. hyalina galeata hybrids and (iii) clones originating from more oligotrophic lakes would be more resistant than clones from more pro-ductive systems. To test these hypotheses, we performed a starvation experiment using D. hyalina and D. hyalina galeata clones from three lakes with contrasting pro-ductivity. To put our results into perspective, we present field data from the most unproductive system used in this study, ultra-oligotrophic Lake Brienz.
M E T H O D
Studied lakes
The Daphnia population of ultra-oligotrophic Lake Brienz (BRZ), Ptot¼ 3.0 mg L21, SRP ¼ 0.9 mg L21,
situated just north of the Swiss Alps, is dominated by D. hyalina, but also hybrids of D. hyalinagaleata and their backcrosses can be found (C. Rellstab, unpub-lished results). In recent years, during the phase of sexual reproduction in autumn, sexual individuals (males, and females with ephippia) represented, on average, 4% of the population (C. Rellstab, unpublished results). However, it is assumed that the main strategy to overwinter is as asexual females, as the lake does not provide ideal conditions for the production and hatch-ing of diapaushatch-ing eggs. Sixty percent of the recent ephippia found in the sediment do not contain eggs, most likely as a result of the low population density
(Rellstab et al., 2007) reducing the chances of mating. Moreover, the high sedimentation rate and the great depth of the lake prevent hatching of the diapausing eggs in most locations (Rellstab et al., 2007).
Mesotrophic Lake Constance (CON), Ptot¼ 13 mg L21
(Stich, 2004), is situated at the border of Switzerland, Germany and Austria. All taxa of the D. hyalina/galeata species complex are present. Daphnia galeata is the predo-minant taxon in summer and produces dormant eggs thereafter. The native species D. hyalina is the most abun-dant in autumn and mainly overwinters asexually. Hybrids also occur throughout the year. Sexual stages in summer or autumn represent an average of 6–8% of the population (Jankowski and Straile, 2004).
Eutrophic Greifensee (GRE), Ptot¼ 63 mg L21
(Keller et al., 2008), is situated in the Swiss lowland. It is dominated by D. hyalina galeata hybrids, but parental D. galeata and D. hyalina can also be found, with the latter being rare (Spaak et al., 2001). Food conditions are sufficient for egg production throughout the winter. During the phase of sexual reproduction, the proportion of sexual stages reaches up to 6% of the population. This can happen up to three times a year, in spring, summer and autumn (P. Spaak, unpublished results).
Field data of Lake Brienz
Qualitative sampling of the Daphnia population in Lake Brienz was performed at least monthly from August 2003 to January 2006 using a single net with a mesh size of 250 mm. Up to 10 net tows from 70-0 m depth were performed to collect enough individuals. On each
sampling date, 80 – 100 asexual females (when present) with a minimum body size of 1 mm (from the top of the eye to the base of the spine) were randomly selected, and the presence and number of eggs was documented. POC concentrations and water temperatures presented are modified from previous studies (Rellstab et al., 2007; Rellstab and Spaak, 2007).
Starvation experiment
We performed a starvation experiment at 58C, using a full factorial design with two-food treatments and two taxa from each of the three lakes described above. Food treatments were: no food (complete absence of food) and low food (adding 0.1 mg C L21 of chemostat-grown Scenedesmus obliquus two times a week). The POC level of the added food was determined by measuring its absorp-tion with a photometer. We chose clones from two taxa from each lake mentioned above: parental D. hyalina (P), and hybrids of D. hyalinagaleata or their backcrosses, hereafter referred to as hybrids (H). Allozyme electro-phoresis was used for taxonomic classification after Keller and Spaak (Keller and Spaak, 2004). Individuals were assayed for the enzyme loci AO and AAT, which are diagnostic markers to distinguish between D. galeata and D. hyalina, and the PGI and PGM loci, which can be further used to identify multi-locus genotypes.
We randomly chose three clones per taxon in each lake, with different multi-locus genotypes (or a different sampling year, if this was not possible, see Table I). An exception was parental D. hyalina of Greifensee, where we had only two clones available. Five replicates were used
Table I: Overview of the clones used in the starvation experiment and their body size on the starting day
of the experiment
Clone Lake Taxon AO AAT PGI PGM Sampling year Size day 1 (mm+++++SE) BRZ H1 Brienz Backcross to D. hyalina ss sf mm ff 2003 0.62 + 0.01 BRZ H2 Brienz Backcross to D. hyalina ss sf ss ff 2003 0.63 + 0.02 BRZ H3 Brienz D. hyalinagaleata F1 sf sf mm ff 2004 0.59 + 0.01 BRZ P1 Brienz Parental D. hyalina ss ss mm ff 2003 0.69 + 0.01 BRZ P2 Brienz Parental D. hyalina ss ss mm f+f+ 2005 0.66 + 0.01 BRZ P3 Brienz Parental D. hyalina ss ss sm ff+ 2005 0.66 + 0.01 CON H1 Constance D. hyalinagaleata F1 sf sf mm sf 2003 0.69 + 0.01 CON H2 Constance D. hyalinagaleata F1 sf+ sf mm ff 2003 0.66 + 0.01 CON H3 Constance D. hyalinagaleata F1 sf+ sf mm mm 2001 0.67 + 0.00 CON P1 Constance Parental D. hyalina ss ss mm ff 2001 0.66 + 0.01 CON P2 Constance Parental D. hyalina ss ss mm ff+ 2001 0.69 + 0.01 CON P3 Constance Parental D. hyalina ss ss mm ff 1998 0.70 + 0.01 GRE H1 Greifensee D. hyalinagaleata F1 sf sf mm sm 2002 0.54 + 0.00 GRE H2 Greifensee D. hyalinagaleata F1 sf sf mf mf 2002 0.64 + 0.01 GRE H3 Greifensee Backcross to D. hyalina ss sf mm ff 2003 0.62 + 0.01 GRE P1 Greifensee Parental D. hyalina ss ss mm ff 2003 0.63 + 0.01 GRE P2 Greifensee Parental D. hyalina ss ss mm sf 2003 0.62 + 0.00 Taxa were determined by allozyme electrophoresis using the two species specific markers AO and AAT. Alleles in AO, AAT, PGI and PGM: s ¼ slow, m ¼ medium, f ¼ fast, f+ ¼ very fast.
for each clone. This resulted in 170 experimental units. Each unit consisted of a single neonate that was placed in a 100 mL jar. Mortality and fecundity (number of juven-iles born) for all animals were recorded at least three times per week. If juveniles were born, they were separated from their mother and put into a no food treatment.
Only second and third clutch juveniles born within 24 h of adult second and third clutch females (originat-ing from one s(originat-ingle mother per clone) were used. These mothers were reared at 128C in filtered (0.45 mm) lake water, fed by adding 1.0 mg C L21 three times a week. Because the age at first reproduction fluctuated broadly between clones, lakes and taxa (16 – 24 days), the start-ing day of the experiment differed among clones. For each clone, initial body size was obtained from 10 juveniles that were not used in the experiment.
To simulate winter conditions, we performed the experiment (including water exchange) in a walk-in growth chamber set at 58C with an 8:16 hour light:dark photoperiod. We used filtered (0.45 mm) water taken from oligotrophic Lake Lucerne in December 2004, in order to have as little dissolved phosphorus as possible and so that none of the clones would have the advan-tage of experiencing water from its lake of origin during the experiment. Water was changed once a week. The experiment lasted from 28 March to 3 August 2005.
Statistical analysis
To evaluate differences in the body size of the neonates on day 1, we performed a three-way ANOVA with lake and taxon as fixed factors, and clone (nested within the lake and taxon) as random factor. Pearson’s r was calculated separately for both food treatments to test whether the average body size of the neonates correlated with the average lifespan and population growth rate r for each clone. We used a four-way ANOVA to test the effect of clone (random factor, nested within lake and taxon), taxon, lake and food concentration (all fixed factors) on lifespan and on r. Analyses of lifespan were based on indi-viduals, whereas analyses of r were done on a clonal level.
The population growth rate r (intrinsic rate of increase) was calculated for each clone that reproduced using the Euler – Lotka equation:
1¼X
n x¼1
lxmxerx
where lx represents the age-specific survivorship of the
clone, mxis the average number of offspring per
surviv-ing individual and x is the age (in days). In clones where no offspring were produced, we calculated an
average population growth rate r by using the following formula: r ¼ (ln(Nx)/ln(N0))/x where Nx; 0.01 (number
of animals at the end of the experiment), N0; 5.01
(number of animals at the start of the experiment) and x the number of days until all animals of the clone had died. The constant value of 0.01 was added to N to make the calculation feasible.
To test whether there was a correlation between the lifespan of the mother and its offspring, Pearson’s r was calculated. All error values given in the text represent the standard error.
R E S U LT S
Field data of Lake Brienz
Lake Brienz is a cold and unproductive system, especially in winter. From 2003 to 2006, average water temperature (Fig. 1A) in the upper 30 m never exceeded 128C (epilimnetic summer temperature reached 15 –
Fig. 1. (A) Average water temperature (black) and average POC concentration (white) of Lake Brienz from August 2003 to January 2006 (upper 30 m). Data modified from Rellstab et al. (2007) and Rellstab and Spaak (2007). (B) Proportion of gravid females (white) and average number of eggs of these females (black) in Lake Brienz from August 2003 to January 2006.
208C in summer), and dropped below 5 – 68C from January to March. Average POC concentrations (Fig. 1A) rarely reached 0.2 mg C L21 in summer and were below 0.1 mg C L21for 5 months in winter (com-plete dataset only for 2004). These data served as a basis for the POC treatments and water temperature used in the experiment. During winters of the study period, Daphnia density dropped below the detection level of the routine sampling program for several con-secutive months, reaching population growth rates of 20.10 to 20.15 days21(Rellstab et al., 2007). However, individuals were always present in the qualitative samples used for this study, as a larger volume of filtered water was analysed. Fecundity data of adult females of Daphnia (Fig. 1B) follow the dynamics of temperature and carbon concentration. Fecundity sharply decreases with the start of winter. In winter 2004, no gravid females could be found for 3 months; in winter 2005, gravid females were rarely found. Fecundity increased again in April in both years.
Starvation experiment: body size of neonates
The body size of neonates differed significantly (P , 0.05) between lakes, taxa and clones (nested within lake and taxon). Lake Constance neonates were the largest, followed by Lake Brienz and Greifensee (Table I). Parental D. hyalina neonates were larger than hybrids. The correlation between body size of neonates and lifespan of Daphnia was close to significance in the no food treatment (n ¼ 17, Pearson’s r ¼ 0.451, P ¼ 0.069), but not in the low food treatment (n ¼ 17, Pearson’s r ¼ 0.252, P ¼ 0.329). Moreover, body size of neonates was significantly correlated to population growth rate r in the no food treatment (n ¼ 17, Pearson’s r ¼ 0.483, P , 0.05), but not in the low food treatment (n ¼ 17, Pearson’s r ¼ 0.226, P ¼ 0.383).
Starvation experiment: lifespan
In both food treatments, the first juveniles died after 4 days. Greatest lifespan was 36 days (individual from CON P3) without food and 126 days (individual from CON H3) under low food conditions. Considering survi-val of the experimental animals over time, irrespective of the taxon and clone, Lake Brienz and Greifensee Daphnia showed high mortality in the no food treatment starting on day 5; all animals finally died within 22 or 13 days, respectively (Fig. 2). Lake Constance clones performed better under no food conditions; the first animal died on day 10. Mortality then increased strongly, but slowed down after day 14. In the low food treatment, a
consistent pattern can be seen for all lakes: high mortality on days 9– 11, followed by low mortality, again a phase of high mortality from days 17–20, again followed by low mortality. Thereafter, Lake Constance and Greifensee animals did not show high mortality until around the period when some animals produced their first clutch (days 66 and 90, respectively).
The results of the four-way ANOVA are presented in Table II and the average lifespan of the different clones are shown in Fig. 3. Food concentration had a highly significant influence on the lifespan of Daphnia: individ-uals without food had much shorter lifespans (12.2 + 0.8 days) compared to those in the low food treatment (43.3 + 4.1 days). Lake Constance animals lived the longest (37.6 + 4.9 days), followed by Greifensee (26.1 + 4.2 days) and Lake Brienz (19.3 + 2.6 days). However, lake origin was not a significant factor in the statistical analysis. In the no food treatment (Lake Constance 17.2 + 1.7 days.Lake Brienz 10.8 + 0.9 days.Greifensee 7.9 + 0.4 days) was the order of survi-val success different than in the low food treatment (Lake Constance 58.0 + 8.2 days.Greifensee 44.4 + 6.6 days.Lake Brienz 27.9 + 4.6 days). Lake Brienz clones clearly showed the smallest reduction in lifespan between the low and no food treatment, compared to the clones of the other two lakes. There was also a significant interaction effect of clone and lake origin. The effect of the taxon and all other interactions were not significant in either analysis.
Starvation experiment: reproductive
success and population growth rater
None of the animals in the no food treatment produced offspring, whereas in the low food treatment a total of 25 neonates were born. Brood size ranged from 1 to 5 (individual from clone GRE P2) eggs. No ephippia were produced in any treatment. Population growth rate r could only be calculated using the Euler – Lotka equation in the clones that reproduced (BRZ P2, CON H1, CON H3, GRE H2, GRE P2). No animals were able to produce a second clutch before they died. As a result of this fact and the low clutch size, r was only slightly positive in clones CON H3 and GRE P2. All other clones showed a negative population growth rate in both food treatments (Fig. 4). All clones except BRZ H3 showed higher population growth rates in the low food treatment (average 20.05 + 0.02 days21) com-pared to conditions without food (20.22 + 0.02 days21). The four-way ANOVA revealed a highly signifi-cant effect of the food treatment and of clone on the population growth rate r of Daphnia (Table II). Moreover, the interaction of LakeFood was significant.
Whereas Greifensee clones had a higher r in the low food treatment (20.02 + 0.03 days21) than Lake Constance and Lake Brienz clones (20.05 + 0.03 days21 and 20.09 + 0.03 days21, respectively), they had a lower r in the no food treatment (20.30 + 0.03 days21) compared to Lake Constance and Lake Brienz clones (20.15 + 0.03 days21and 20.20 + 0.03 days21,
respectively). All other factors and their interactions were not significant.
Age at first reproduction differed greatly between lakes. It took an average of 66 days for the BRZ and GRE clones to produce neonates, while CON clones needed around 90 days.
Fig. 2. Survival curves of Daphnia originating from three different lakes under no food (top) and low food (bottom) conditions.
Table II: Results of the four-way ANOVA testing the effect of clone, lake, taxon and food treatment on
the lifespan (left) and population growth rate r (right) of Daphnia
Dependent variable Lifespan Population growth rate r
Effect df MS FF PP df MS FF PP Intercept 1 129284.9 57.0 ,0.001 1 0.601 74.537 ,0.001 Clone (Lake Taxon) 11 2266.2 1.78 0.192 11 0.008 3.383 0.027 Lake 2 5063.3 2.2 0.153 2 0.012 1.440 0.278 Taxon 1 1030.9 0.5 0.514 1 0.004 0.441 0.520 Food 1 41796.6 31.6 ,0.001 1 0.217 91.225 ,0.001 Clone (Lake Taxon) Food 11 1320.3 2.8 0.002 11 0.002
Lake Taxon 2 708.7 0.3 0.738 2 0.001 0.069 0.934 Lake Food 2 2392.5 1.8 0.209 2 0.025 10.887 0.002 Taxon Food 1 872.7 0.7 0.433 1 0.001 0.327 0.578 Lake Taxon Food 2 301.5 0.2 0.800 2 0.001 0.223 0.803
Error 136 468.4 0
Starvation experiment: offspring survival Offspring produced during the experiment were put into a no food treatment. Lifespans of these juveniles
ranged between 2 and 25 days (average lifespan per brood between 2.0 and 23.5 days). Juveniles from Lake Constance lived the longest (16.3 + 2.3 days), followed
Fig. 3. Average lifespan of the different clones in the starvation experiment under no food and low food conditions. Clones belong to three different lakes (left to right) and two different taxa (top and bottom). Error bars represent standard error.
Fig. 4. Population growth rate r of the different clones in the starvation experiment under no food and low food conditions. Clones belong to three different lakes (left to right) and two different taxa (top and bottom).
by Lake Brienz (6.0 + 0.0 days) and Greifensee (4.9 + 0.5 days). There was a significant and strong correlation (n = 25, Pearson’s r ¼ 0.742, P , 0.001) between the lifespan of the mother and the lifespan of her offspring.
D I S C U S S I O N
Our results show that Daphnia can be resistant to star-vation for long periods at low temperature, which indicates that they can successfully overwinter as parthe-nogenetic animals in ultra-oligotrophic lakes. The taxon and lake from which the daphnids originate did not sig-nificantly influence the survival under winter conditions, but regarding population growth rate, the interaction between lake origin and food concentration was signifi-cant. Reproduction took place only under low food conditions, but was in the majority of cases not high enough to result in a positive population growth rate. The lifespan of the neonates that hatched during the experiment was significantly correlated to the lifespan of their mothers.
Despite significant differences in size at birth at the start of the experiment, we found no significant corre-lation between the body size and the lifespan. However, larger neonates had a significantly higher population growth rate r in the no food treatment. Body size can affect the resistance of neonates to starvation (Tessier and Consolatti, 1989), following the predictions of the size-efficiency hypothesis (Brooks and Dodson, 1965). Another possible reason is that larger neonates contain more energy reserves, as discussed in Tessier and Consolatti (Tessier and Consolatti, 1989). In contrast, Glazier (Glazier, 1992) found no correlation of the body size of neonates and the ability to survive starvation in D. magna.
Daphnia in our experiment showed an average lifespan of 12 days (maximum 36 days) under no food conditions and 43 days (maximum 126 days) under low food con-ditions (Figs 2 and 3). This is higher than in other star-vation studies, where a lifespan under complete absence of food was on average 6 – 9 days for juvenile daphnids (Gliwicz, 1991; Gliwicz and Guisande, 1992; Perrin et al., 1992) and 5 – 11 days for adult daphnids (Lemcke and Lampert, 1975; Threlkeld, 1976; Elendt, 1989) of various Daphnia species. Only Epp (Epp, 1996) reported lifespans of 28 days and more, using neonates of D. pulicaria in the complete absence of food. All of these other experiments were performed at 15 – 208C, so the low water temperature in our experiment (58C) is most likely responsible for the differences in lifespans. As we used filtered lake water (as most of the studies mentioned in this manuscript), it is also possible that
the daphnids in the no food treatment had very low amounts of organic matter other than algae available, e.g. bacteria, which are known to be a potential food source (Kankaala, 1988). The lifespans at low food treat-ments, which represent realistic conditions for an ultra-oligotrophic lake in winter, show that it is possible for at least some Daphnia clones to overwinter successfully as parthenogenetic animals.
The food concentration had a significant effect on the lifespan of daphnids, whereas lake and taxon did not play a substantial role. We also found a significant interaction of clone origin and food concentration, meaning that clones showed different patterns in response to different food concentrations. Inter-clonal variation in resistance to starvation was also shown by Epp (Epp, 1996). Surprisingly, it was not the clones from ultra-oligotrophic Lake Brienz, but the clones from mesotrophic Lake Constance that had the highest lifespans under no food conditions. Lake Brienz clones performed slightly better than clones from eutrophic Greifensee. Maternal effects, which can be environ-mental or genetic (Ra¨sa¨nen and Kruuk, 2007), are most likely the cause of the variation in lifespan. These play an important role in the resistance to starvation because mothers transfer energy reserves like lipids to their off-spring (Tessier et al., 1983; Cowgill et al., 1984), and produce neonates of different sizes (Gliwicz and Guisande, 1992). Especially in our no food treatment, neonates depended only on the energy reserves received from their mothers. Prior to our experiment intra- and inter-clonal differences due to maternal effects resulting from different environmental conditions were elimi-nated, because several generations were treated exactly the same before their juveniles were exposed to the treatments. Still, as the mothers and grandmothers of our experimental animals were reared at the same rela-tively high food concentration, our results could par-tially represent the efficiency at which the mothers feed at high food concentration and transfer lipids to their juveniles. The strong correlation between the survival of Daphnia that reproduced in the low food treatment and the survival of its offspring in the no food treatment, and the significant effect of neonate body size on popu-lation growth rate, further support the importance of (genetic) maternal effects.
By calculating the population growth rate r for each clone in both food treatments, we obtained a measure based not only on survival, but also on reproduction. Only two clones managed to produce enough offspring to have a positive population growth rate, all others were negative. With daily death rates of up to 36% under no food and up to 18% under low food, our results are similar to or slightly lower than death rates found in other
studies (Lemcke and Lampert, 1975; Lynch, 1989; Rohrlack et al., 1999). Our ANOVA showed a significant effect of food treatment and its interaction with lake origin on population growth rate r. Moreover, there was a signifi-cant inter-clonal variation. There are remarkable differ-ences in the outcome of the statistical analyses concerning lifespan and population growth rate. Because our popu-lation growth rates are also mostly derived from lifespan (we could calculate r using the Euler–Lotka equation in only five clones), these differences must come from the fact that the temporal distribution of deaths is not incor-porated. It represents the average growth rate from the start of the experiment until the last individual of the clone had died.
Assuming that the variation in the lifespan and growth rate is determined mainly by genetic maternal effects, we could use the difference of the response (reaction norm) to the two food treatments as a measure of resistance to star-vation. Our ANOVAs showed a significant LakeFood interaction for r, meaning that the reduction in population growth rate is different among lakes when comparing the low and no food treatment. If we look at lifespan, Daphnia clones from Lake Brienz show a remarkably smaller differ-ence between the two food treatments (16.4 days) com-pared to those from Greifensee (36.5 days) or Lake Constance (40.8 days). The reduction of population growth rate r is much higher in Greifensee clones (0.28 days21) than in clones from Lake Constance (0.10 days21) or Lake Brienz (0.10 days21). This is an indication that, when looking at lifespan, Daphnia from Lake Brienz perform best, and when considering population growth rate, Daphnia from Greifensee are performing worst. In fact, a three-way ANOVA with Clone (random factor, nested in LakeTaxon), and Lake and Taxon as fixed factors shows a significant lake effect (df ¼ 2, F ¼ 10.886, P , 0.05) on the difference in population growth rate (but not lifespan) between the two food treatments. This agrees with our original prediction that clones from an ultra-oligotrophic lake show higher fitness under starvation.
Taxa of the D. hyalina/galeata species complex differ in many life history traits, including migration behaviour (Weider and Stich, 1992), timing and extent of sexual reproduction (Jankowski and Straile, 2004), genetic diversity (Jankowski and Straile, 2004; Keller and Spaak, 2004), parasite prevalence (Wolinska et al., 2006), predation sensitivity (Spaak et al., 2000) and habitat pre-ference (Keller et al., 2008). We could not verify our hypothesis that parental D. hyalina, which are the domi-nant taxon in oligotrophic lakes (Keller et al., 2008), would perform better under starvation conditions than hybrids. Why could we not find an effect? Several reasons might play a role. First, the taxonomic compo-sition present in these lakes might not be determined by
the ability of a taxon to resist starvation in winter. Other selection factors, such as feeding efficiency under better food conditions or overwintering success in diapause could be more important. Secondly, the clones used in this experiment were not collected specifically at the end of winter when we would expect the better adapted clones to be more abundant. Thirdly, using two species specific markers can lead to misidentification of taxa due to backcrossing of hybrids with the parental taxa. A part of parental D. hyalina or D. hyalinagaleata clones could actually represent backcrosses to D. hyalina as is most likely the case in Greifensee for D. hyalina (Keller and Spaak, 2004) and in Lake Brienz for D. hyalina galeata (C. Rellstab, unpublished results).
Field data from Lake Brienz presented here and in a previous study (Rellstab et al., 2007) show that winter is the major bottleneck for a Daphnia population in an ultra-oligotrophic lake. Temperature and food concentration are at very low levels, resulting in low fecundity and nega-tive population growth rates (similar to the ones found in our experiment) for several consecutive months. However, asexual females are always present throughout the winter and represent an important part of the starting population in spring, when conditions ameliorate. Due to the low population density (reducing the chances of mating), the large depth and the high sedimentation rate (reducing hatching success), it is unlikely that overwintering by sexu-ally produced resting eggs is a successful strategy. Note that under natural conditions, the effect of predation by fish on the asexually overwintering population should not be underestimated (Jeppesen et al., 2004), although in Lake Brienz predation pressure is low (Mu¨ller et al., 2007b). Even low predation pressure could have a strong impact due to the low population growth rates found in our experiment.
Since the growth rates in our experiment are mostly negative, it appears that it is the individual ability to withstand starvation in winter, rather than significant population growth, that permits clones to be present in spring. Thus, the probability of a female’s offspring sur-viving the period of starvation must, at least in some years, exceed their probability of surviving through winter as diapausing eggs and then hatching success-fully. As the trophic state of the lake changes, previous high-fitness life history strategies might become poorly adaptive. The clonal variation we showed may provide the basis for evolutionary response by Daphnia, but will depend upon the strength of selection and the available genetic variation (Kinnison and Hairston, 2007). In Lake Brienz, oligotrophication has led to substantial declines in fish stocks (Mu¨ller et al., 2007a, 2007b). Understanding how their major food resource survives and adapts to the stressful conditions of generally
decreasing food conditions will be critical in evaluating how to manage the balance between improved water quality and the needs of the fishery.
In conclusion, we have shown that the response of Daphnia to realistic winter conditions is highly variable among clones, and is influenced mainly by food con-centration and by the lake origin. Taxon did not play a significant role. The absolute performance of the differ-ent clones did not match our expectation that clones from unproductive systems would perform better under their local conditions. Still, the significant interaction of lake origin and food concentration indicates that these clones perform better if we look at the difference of the response to the two food conditions. Moreover, consid-ering results found in the literature and from our experiment, there is evidence that maternal effects play a major role, especially when neonates are exposed to the complete absence of food. Our results show that it is possible for Daphnia clones to withstand long periods of low food conditions (typical for ultra-oligotrophic systems in winter), and short periods of complete star-vation. Reproduction increases population growth rate until conditions ameliorate, but does not necessarily result in positive growth, since clutch sizes are too small and egg development time is too long.
AC K N O W L E D G E M E N T S
We thank Esther Keller for assistance during the experi-ment, Christine Dambone for maintenance of the che-mostat, Justyna Wolinska for support in planning and analysing, Whit Hairston for linguistic help, and Jukka Jokela, Nelson Hairston, Luc De Meester, Kirstin Kopp and two anonymous reviewers for useful comments on the manuscript.
F U N D I N G
The present study is part of an interdisciplinary research project investigating the ecological impacts of anthropogenic changes in Lake Brienz and its catch-ment. The study was funded by the government of the Canton of Bern, KWO Grimselstrom, Federal Office for the Environment (FOEN), Eawag and the local communities.
R E F E R E N C E S
Bottrell, H. H. (1975) Relationship between temperature and duration of egg development in some epiphytic Cladocera and Copepoda
from River Thames, Reading, with a discussion of temperature functions. Oecologia,18, 63 – 84.
Bradley, M. C., Perrin, N. and Calow, P. (1991) Energy allocation in the cladoceran Daphnia magna Straus, under starvation and refeed-ing. Oecologia,86, 414 – 418.
Brooks, J. L. and Dodson, S. I. (1965) Predation, body size, and com-position of plankton. Science,150, 28 – 35.
Burns, C. W. (1968) Relationship between body size of filter-feeding Cladocera and maximum size of particle ingested. Limnol. Oceanogr., 13, 675 – 678.
Cowgill, U. M., Williams, D. M. and Esquivell, J. B. (1984) Effects of maternal nutrition on fat content and longevity of neonates of Daphnia magna. J. Crust. Biol.,4, 173 – 190.
De Meester, L., Vanoverbeke, J., De Gelas, K. et al. (2006) Genetic structure of cyclic parthenogenetic zooplankton populations—a conceptual framework. Arch. Hydrobiol.,167, 217 – 244.
Elendt, B. P. (1989) Effects of starvation on growth, reproduction, sur-vival and biochemical composition of Daphnia magna. Arch. Hydrobiol., 116, 415 – 433.
Epp, G. T. (1996) Clonal variation in the survival and reproduction of Daphnia pulicaria under low-food stress. Freshwater Biol.,35, 1 – 10. Glazier, D. S. (1992) Effects of food, genotype, and maternal size and
age on offspring investment in Daphnia magna. Ecology,73, 910 – 926. Gliwicz, M. Z. (1990) Food thresholds and body size in cladocerans.
Nature,343, 638 – 640.
Gliwicz, M. Z. (1991) Food thresholds, resistance to starvation, and cla-doceran body size. Verh. Int. Ver. Theor. Angew. Limnol.,24, 2795–2798. Gliwicz, M. Z. and Guisande, C. (1992) Family planning in Daphnia—
resistance to starvation in offspring born to mothers grown at differ-ent food levels. Oecologia,91, 463 – 467.
Gliwicz, M. Z., Slusarczyk, A. and Slusarczyk, M. (2001) Life history synchronization in a long-lifespan single-cohort Daphnia population in a fishless alpine lake. Oecologia,128, 368 – 378.
Goss, L. B. and Bunting, L. D. (1983) Daphnia development and repro-duction: responses to temperature. J. Therm. Biol.,4, 375 – 380. Jankowski, T. and Straile, D. (2004) Allochronic differentiation among
Daphnia species, hybrids and backcrosses: the importance of sexual reproduction for population dynamics and genetic architecture. J. Evol. Biol.,17, 312 – 321.
Jeppesen, E., Jensen, J. P., Sondergaard, M. et al. (2004) Impact of fish predation on cladoceran body weight distribution and zooplankton grazing in lakes during winter. Freshwater Biol.,49, 432 – 447. Kankaala, P. (1988) The relative importance of algae and bacteria as
food for Daphnia longispina (Cladocera) in a polyhumic lake. Freshwater Biol.,19, 285 – 296.
Keller, B. and Spaak, P. (2004) Nonrandom sexual reproduction and diapausing egg production in a Daphnia hybrid species complex. Limnol. Oceanogr.,49, 1393– 1400.
Keller, B., Wolinska, J., Tellenbach, C. et al. (2007) Reproductive iso-lation keeps hybridizing Daphnia species distinct. Limnol. Oceanogr., 52, 984 – 991.
Keller, B., Wolinska, J., Manca, M. et al. (2008) Spatial, environ-mental, and anthropogenic effect on the taxa composition of hybri-dizing Daphnia. Phil. Trans. R. Soc. B,363, 2943– 2952.
Kinnison, M. T. and Hairston, N. G. (2007) Eco-evolutionary conser-vation biology: contemporary evolution and the dynamics of persist-ence. Funct. Ecol.,21, 444 – 454.
Lampert, W. (1977) Studies on the carbon balance of Daphnia pulex as related to environmental conditions-IV. Determination of the “threshold” concentration as a factor controlling the abundance of zooplankton species. Arch. Hydrobiol. Suppl.,48, 361 – 368.
Lampert, W. (1978) A field study on the dependence of the fecundity of Daphnia spec. on food concentration. Oecologia,36, 363 – 369. Lemcke, H. W. and Lampert, W. (1975) Vera¨nderungen im Gewicht
und der chemischen Zusammensetzung von Daphnia pulex im Hunger. Arch. Hydrobiol.,48, 108 – 137.
Lynch, M. (1989) The life history consequences of resource depression in Daphnia pulex. Ecology,70, 246 – 256.
Lynch, M. and Ennis, R. (1983) Resource availability, maternal effects, and longevity. Exp. Gerontol.,18, 147 – 165.
Lynch, M., Weider, L. J. and Lampert, W. (1986) Measurement of the carbon balance in Daphnia. Limnol. Oceanogr.,31, 17 – 33.
McCauley, E. M. W. W. and Nisbet, R. M. (1991) Growth reproduc-tion, and mortality of Daphnia pulex Leydig: life at low food. Funct. Ecol.,4, 505 – 514.
Mort, M. A. and Wolf, H. G. (1986) The genetic structure of large lake Daphnia populations. Evolution,40, 756 – 766.
Mousseau, T. A. and Fox, C. W. (1998) The adaptive significance of maternal effects. Trends Ecol. Evol.,13, 403 – 407.
Mu¨ller, B., Finger, D., Sturm, M. et al. (2007a) Present and past bio-available phosphorus budget in the ultra-oligotrophic Lake Brienz. Aquat. Sci.,69, 227 – 239.
Mu¨ller, R., Breitenstein, M., Bia, M. et al. (2007b) Bottom-up control of whitefish populations in ultra-oligotrophic Lake Brienz. Aquat. Sci.,69, 271 – 288.
Pearl, R. (1928) The Rate of Living. Knopf, New York, 185 pp. Perrin, N., Baird, D. J. and Calow, P. (1992) Resource allocation,
population dynamics and fitness: Some experiments with Daphnia magna Straus. Arch. Hydrobiol.,123, 431 – 449.
Plath, K. (1998) Adaptive feeding behavior of Daphnia magna in response to short-term starvation. Limnol. Oceanogr.,43, 593 – 599. Ra¨sa¨nen, K. and Kruuk, L. E. B. (2007) Maternal effects and
evol-ution at ecological time-scales. Funct. Ecol.,21, 408 – 421.
Rellstab, C. and Spaak, P. (2007) Starving with a full gut? Effect of suspended particles on the fitness of Daphnia hyalina. Hydrobiologia, 594, 131 – 139.
Rellstab, C., Maurer, V., Zeh, M. et al. (2007) Temporary collapse of the Daphnia population in turbid and ultra-oligotrophic Lake Brienz. Aquat. Sci.,69, 257 – 270.
Rohrlack, T., Dittmann, E., Henning, M. et al. (1999) Role of micro-cystins in poisoning and food ingestion inhibition of Daphnia galeata caused by the cyanobacterium Microcystis aeruginosa. Appl. Environ. Microbiol.,65, 737 – 739.
Schwenk, K. and Spaak, P. (1995) Evolutionary and ecological conse-quences of interspecific hybridization in cladocerans. Experientia,51, 465 – 481.
Spaak, P., Vanoverbeke, J. and Boersma, M. (2000) Predator induced life history changes and the coexistence of five taxa in a Daphnia species complex. Oikos,89, 164 – 174.
Spaak, P., Eggenschwiler, L. and Bu¨rgi, H. (2001) Genetic variation and clonal differentiation in the Daphnia population of the Greifensee, a pre-alpine Swiss lake. Verh. Int. Ver. Theor. Angew. Limnol.,27, 1919– 1923.
Stich, H. B. (2004) Back again: the reappearance of Diaphanosoma bra-chyurum in Lake Constance. Arch. Hydrobiol. Beih. Ergebn. Limnol.,159, 423 – 431.
Tessier, A. J. and Consolatti, N. L. (1989) Variation in offspring size in Daphnia and consequences for individual fitness. Oikos, 56, 269 – 276.
Tessier, A. J., Henry, L. L. and Goulden, C. E. (1983) Starvation in Daphnia: energy reserves and reproductive allocation. Limnol. Oceanogr.,28, 667 – 676.
Threlkeld, S. T. (1976) Starvation and the size structure of zooplank-ton communities. Freshwater Biol.,6, 489 – 496.
Vijverberg, J. (1980) Effect of temperature in laboratory studies on development and growth of Cladocera and Copepoda from Tjeukemeer, The Netherlands. Freshwater Biol.,10, 317 – 340. Weider, L. J. and Stich, H. B. (1992) Spatial and temporal
heterogen-eity of Daphnia in Lake Constance; intra- and interspecific compari-sons. Limnol. Oceanogr.,37, 1327– 1334.
Wolinska, J., Bittner, K., Ebert, D. et al. (2006) The coexistence of hybrid and parental Daphnia: the role of parasites. Proc. R. Soc. Lond. B Biol. Sci.,273, 1977– 1983.
Yurista, P. M. (1999) Temperature-dependent energy budget of an Arctic Cladoceran. Daphnia middendorffiana. Freshwater Biol., 42, 21 – 34.
